Mathias Brandewinder on .NET, F#, VSTO and Excel development, and quantitative analysis / machine learning.
by Mathias 25. March 2013 10:33

My trajectory through “Machine Learning in Action” is becoming more unpredictable as we go – this time, rather than completing our last episode on K-means clustering (we’ll get back to it later), I’ll make another jump directly to Chapter 14, which is dedicated to Singular Value Decomposition, and convert the example from Python to F#.

The chapter illustrates how Singular Value Decomposition (or SVD in short) can be used to build a collaborative recommendation engine. We will follow the chapter pretty closely: today we will focus on the mechanics of using SVD in F# – and leave the recommendation part to our next installment.

As usual, the code is on GitHub.

Until this point, I have avoided using a Linear Algebra library, because the algorithms we discussed so far involved lightweight, row-centric operations, which didn’t warrant taking such a dependency. SVD is one of these cases where using an established library is a good idea, if only because implementing it yourself would not be trivial. So let’s create a new script file (Chapter14.fsx), add a reference to Math.NET Numerics for F# to our project via NuGet, and reference it in our script:

#r @"..\..\MachineLearningInAction\packages\MathNet.Numerics.2.4.0\lib\net40\MathNet.Numerics.dll"
#r @"..\..\MachineLearningInAction\packages\MathNet.Numerics.FSharp.2.4.0\lib\net40\MathNet.Numerics.FSharp.dll"

open MathNet.Numerics.LinearAlgebra
open MathNet.Numerics.LinearAlgebra.Double

Now that we have our tools, let’s start working our example. Imagine that we are running a website, where our users can rate dishes, from 1 (horrendous) to 5 (delightful). Our data would look something along these lines:

type Rating = { UserId: int; DishId: int; Rating: int }

// Our existing "ratings database"
let ratings = [
    { UserId = 0; DishId = 0; Rating = 2 };
    { UserId = 0; DishId = 3; Rating = 4 };
    ... omitted for brevity ...
    { UserId = 10; DishId = 8; Rating = 4 };
    { UserId = 10; DishId = 9; Rating = 5 } ]

Our goal will be to provide recommendations to User for Dishes they haven’t tasted yet, based on their ratings and what other users are saying.

Our first step will be to represent this as a Matrix, where each Row is a User, each Column a Dish, and the corresponding cell is the User Rating for that Dish. Note that not every Dish has been rated by every User – we will represent missing ratings as zeroes in our matrix:

let rows = 11
let cols = 11
let data = DenseMatrix(rows, cols)
|> List.iter (fun rating -> 
       data.[rating.UserId, rating.DishId] <- (float)rating.Rating)

We initialize our 11 x 11 matrix, which creates a zero-filled matrix, and then map our user ratings to each “cell”. Because we constructed our example that way, our UserIds go from 0 to 10, and DishIds from 0 to 10, so we can map them respectively to Rows and Columns.

Note: while this sounded like a perfect case to use a Sparse Matrix, I chose to go first with a DenseMatrix, which is more standard. I may look at whether there is a benefit to going sparse later.

Note: our matrix happens to be square, but this isn’t a requirement.

Note: I will happily follow along the book author and replace unknown ratings by zero, because it’s very convenient. I don’t fully get how this is justified, but it seems to work, so I’ll temporarily suspend disbelief and play along.

At that point, we have our data matrix ready. Before going any further, let’s write a quick utility function, to “pretty-render” matrices:

let printNumber v = 
    if v < 0. 
    then printf "%.2f " v 
    else printf " %.2f " v
// Display a Matrix in a "pretty" format
let pretty matrix = 
    Matrix.iteri (fun row col value ->
        if col = 0 then printfn "" else ignore ()
        printNumber value) matrix
    printfn ""

We iterate over each row and column, start a newline every time we hit column 0, and print every value, nicely formatted with 2 digits after the decimal.

In passing, note the F#-friendly Matrix.iteri syntax – the good people at Math.NET do support F#, and MathNet.Numerics.FSharp.dll contains handy helpers, which allow for a much more functional usage of the library. Thanks, guys!

Let’s see how our data matrix looks like:

printfn "Original data matrix"
pretty data

… which produces the following output in FSI:

Original data matrix

2.00  0.00  0.00  4.00  4.00  0.00  0.00  0.00  0.00  0.00  0.00
0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  5.00
4.00  0.00  0.00  0.00  0.00  0.00  0.00  1.00  0.00  0.00  0.00
3.00  3.00  4.00  0.00  3.00  0.00  0.00  2.00  2.00  0.00  0.00
5.00  5.00  5.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00
0.00  0.00  0.00  0.00  0.00  0.00  5.00  0.00  0.00  5.00  0.00
4.00  0.00  4.00  0.00  0.00  0.00  0.00  0.00  0.00  0.00  5.00
0.00  0.00  0.00  0.00  0.00  4.00  0.00  0.00  0.00  0.00  4.00
0.00  0.00  0.00  0.00  0.00  0.00  5.00  0.00  0.00  0.00  0.00
0.00  0.00  0.00  3.00  0.00  0.00  0.00  0.00  4.00  5.00  0.00
1.00  1.00  2.00  1.00  1.00  2.00  1.00  0.00  4.00  5.00  0.00

We seem to be in business.


by Mathias 31. October 2012 13:43

This post is to be filed in the “useless but fun” category. A friend of mine was doing some Hadoopy stuff a few days ago, experimenting with rather large sparse matrices and their products. Long story short, we ended up wondering how sparse the product of 2 sparse matrices should be.

A sparse matrix is a matrix which is primarily filled with zeros. The product of two matrices involves computing the dot product of each row of the first one, with each column of the second one. There is clearly a relationship between how dense the two matrices are, and how dense the result should be. As an obvious illustration, if we consider 2 matrices populated only with zeroes – as sparse as it gets – their product is obviously 0. Conversely, two matrices populated only with ones – as dense as it gets – will result in a “completely dense” matrix. But… what happens in between?

Should I expect the product of 2 sparse matrices to be more, or less dense than the original matrices? And does it depend on how large the matrices are? What would be your guess?


by Mathias 7. April 2012 10:50

For no clear reason, I got interested in Convex Hull algorithms, and decided to see how it would look in F#. First, if you wonder what a Convex Hull is, imagine that you have a set of points in a plane – say, a board – and that you planted a thumbtack on each point. Now take an elastic band, stretch it, and wrap it around the thumbtacks. The elastic band will cling to the outermost tacks, leaving some tacks untouched. The convex hull is the set of tacks that are in contact with the elastic band; it is convex, because if you take any pair of points from the original set, the segment connecting them remains inside the hull.

The picture below illustrates the idea - the blue thumbtacks define the Convex Hull; all the yellow tacks are included within the elastic band, without touching it.


There are a few algorithms around to identify the Convex Hull of a set of points in 2 dimensions; I decided to go with Andrew’s monotone chain, because of its simplicity.

The insight of the algorithm is to observe that if you start from the leftmost tack, and follow the elastic downwards, the elastic turns only clockwise, until it reaches the rightmost tack. Similarly, starting from the right upwards, only clockwise turns happen, until the rightmost tack is reached. Given that the left- and right-most tacks belong to the convex hull, the algorithm constructs the upper and lower part of the hull by progressively constructing sequences that contain only clockwise turns.


by Mathias 21. August 2011 16:15

Let’s take a last stab at our beer-delivery problem. We tried out a Sieve, we used the Microsoft Solver – time for some recursion.

How can we organize our recursion?

If we had only 1 type of beer pack, say, 7-packs, the best way to supply n bottles of beer is to supply the closest integer greater than n/7, that is, $$\lceil {n \over 7} \rceil$$

If we had 7-packs and 13-packs, we need to consider multiple possibilities. We can select from 0 to the ceiling of n/7 7-packs, and, now that we have only one type of case pack left, apply the same calculation as previously to the remaining bottles we need to supply – and select the best of the combinations, that is, the combination of beer packs closest to the target.

If we had even more types of beer packs available, we would proceed the same way, by trying out the possible quantities for the first pack, and given the first, for the second, and so on until we reach the last type of pack – which is pretty much the outline of a recursive algorithm.


by Mathias 14. August 2011 13:41

In our last post, we looked at a Sieve-like algorithm to help a Brewery find how closely they can match the number of beer bottles their thirsty customers desire, using only 7-packs and 13-packs of delicious beer; in less appetizing but more precise terms, we are trying to solve the following problem:

Suppose that you are given a list of integers, and a target integer. Your goal is to find the closest value that is greater or equal to the target, by combining integers (“packs”) from the list (only positive combinations are allowed). For instance, given 3 and 5, the closest you can get to 16 would be 5 x 2 + 3 x 2 = 16, and the closest to 17 would be 3 x 6 = 18.

The Sieve solution is pretty effective, but has some limitations. Today, we’ll take another approach: leveraging the Microsoft Solver Foundation.

The beauty of the Solver is that it allows you to focus on what you want to achieve, rather than on how to achieve it. As long as you can define clearly what your goal, your decision variables and your constraints are, you can leave it to the Solver engine to figure out what the best way to achieve that goal is, by searching the best values for the Decision variables you defined.

So what are we trying to achieve here? Our goal, in Solver terms, is to minimize the extra number of bottles shipped, under the constraint that the number of bottles shipped is greater than the requested target number. Our Decision variables are the number of units of each Beer Pack we will ship, with a constraint that Decisions must be integer (we cannot ship half-packs), and positive.

Let’s add a reference to the Solver in our project (details here), and see how this looks like in code:



Comment RSS